Where we are
You have a ramp and a written bar. Now the recurring work: the weekly conversation in which capability actually changes. Start from the distinction that makes coaching hard for competent people. Fixing transfers the answer; coaching transfers the ability to find answers. Fixing is genuinely correct sometimes - when the building is on fire, tell people where the exit is - but it has a hidden cost, which is that a team that is always fixed never develops and eventually cannot function without you. The awkward truth for anyone promoted for being excellent at the work: your strongest instinct is now your main obstacle, because you can see the answer in ten seconds and coaching means not saying it for twenty minutes.
GROW: goal, reality, options, will 7 min live
Whitmore's GROW model has lasted because it solves the actual problem with coaching conversations, which is that they drift. Four stages, in order, and the discipline is finishing each one before moving on - most failed coaching conversations jump from a half-stated problem straight to the manager's preferred solution.
LiveThe three questions that carry most of the work3 min▶
- "What have you tried?" Ask it before offering anything. Half the time they have already tried your idea, and advising it anyway signals that you were not listening. The other half, the answer tells you exactly where the gap is.
- "And what else?" First options are usually the obvious ones. The third or fourth is where the thinking happens. This question costs four words and does more than any framework.
- "What would you do if I were on holiday?" The unblocker for someone who has learned to bring you decisions. It reveals both that they usually have a view and that they have been trained not to trust it.
And one thing to say out loud when you do decide to fix rather than coach: "I am going to just tell you this one because of the deadline - but let us come back to how you would work it out." That keeps fixing honest instead of accidentally becoming the default.
Oversight follows the task, not the title 5 min live
Grove's task-relevant maturity is the most useful idea in management and the most consistently misapplied. Maturity is not a property of a person - it is a property of this person on this task. The same individual can need close direction on one thing and complete autonomy on another in the same week.
LiveCoaching, feedback, mentoring, sponsorship - four different things3 min▶
These get used interchangeably, and the confusion causes real problems - most commonly a manager who thinks they are coaching when they are giving feedback, or an organisation that offers mentoring when the gap is sponsorship.
- Coaching - you ask, they think, they decide. Builds capability. Forward-looking. Your job is questions.
- Feedback - you tell them what you observed and its effect. Corrects a specific behaviour. Backward-looking about an event, forward-looking about a change. Your job is precision.
- Mentoring - you share experience and perspective, usually broader than the current job. Optional, and often better from someone outside their reporting line.
- Sponsorship - you spend your credibility on them: putting their name forward, giving them the visible assignment, arguing their case in rooms they are not in. This is the one that changes careers, and the one your team cannot ask you for directly.
A team member who needs sponsorship and is offered coaching will feel supported and go nowhere - which is exactly the pattern the leader track's promotion session finds in the data.
The manager who coached instead of sponsoring. A Cartwheel team lead ran excellent weekly coaching with a strong analyst for two years. The analyst grew visibly and was passed over for promotion twice, because nobody in the room where levelling happened had ever seen her work or heard her name argued for. Her manager had done the developmental half of the job perfectly and none of the political half. Coaching builds the capability; sponsorship makes it count. Both are your job.
Run a GROW conversation on a real Cartwheel problem ★ 12 min · everyone tries
A Cartwheel analyst says: "The buyers keep ignoring my forecast." The fixing answer is available immediately and is a trap. Run the four stages instead, in pairs, and notice how much you learn that the fix would have skipped.
Goal. "What would be different if this were working?" Note whether their goal is what you assumed - it often is not. Here it turns out to be credibility, not compliance.
Reality. "What happens now, specifically? Which buyers, which decisions, what do they use instead?" Facts before theories. This is where you discover the forecast arrives after the buying meeting.
Options. "What could you do?" Then "and what else?" twice. Expect three to five options, of which the timing one was not their first thought.
Will. "What will you do, by when, and what do you need from me?" Get a specific commitment and name your own part in it.
| Stage | Fixing version | Coaching version |
|---|---|---|
| Opening | "Send it earlier and put it in the channel." | "What would be different if this were working?" |
| What surfaced | Nothing - the manager's guess was the whole input | The forecast lands after the buying meeting, so it cannot be used at all |
| Who owns it | The manager. They will be asked again next week | The analyst, who chose the change |
| Time spent | Two minutes | Fifteen minutes, once |
| Capability after | Unchanged | They now know how to diagnose a "nobody listens to me" problem |
Sort your own team by task and maturity ★ 8 min · do your own
Not by person - by person-and-task. The grid usually reveals both failure modes running simultaneously on the same team.
List each person and their two or three current pieces of work separately.
Rate maturity per line: new to this, done it a few times, owns it.
Write the oversight you are actually giving each line, honestly.
Find your mismatches, and check specifically for the senior-person-new-task trap.
Both failures on one team, same week. A Cartwheel lead was reviewing every line of her most experienced analyst's weekly report - a task he had owned for three years, which he read as distrust and mentioned in his exit interview. In the same week she had handed the same analyst the new supplier-negotiation work with no support at all, on the grounds that he was senior. He was highly mature on one task and completely new to the other, and she had precisely inverted the support each needed. Rate the task, not the person.
Try it yourself - this week ◐ 30-45 min total
- Run one full GROW conversation without offering a single solution. Notice how uncomfortable the silences are, and hold them anyway.
- Ask "and what else?" twice in one conversation and write down what came out third.
- Build the person-by-task maturity grid for your team and fix one mismatch.
- Check for the senior-person-new-task trap specifically. It is the most common and most expensive one.
- Name one person who needs sponsorship rather than coaching, and do one concrete sponsoring act for them this month.
- Next time you choose to fix rather than coach, say so out loud and name when you will come back to it.
Sources covered (cited)
Taught from the coaching canon. This session covers:
Three questions before you go 🎯 ◐ 90 seconds
1 · What is the most common way a GROW conversation collapses?
Twelve seconds, correct-sounding advice, zero capability transferred - and they will be back next week. Worse, the fix is often wrong, because you skipped the Reality stage where the actual cause was hiding.
2 · Task-relevant maturity is a property of...
The classic trap is giving your most senior person the least support on a task that is brand new to them. Rate the task, not the person - and watch for both failures at once: micromanaging high maturity reads as distrust, abandoning low maturity produces the month-nine performance case.
3 · A strong analyst is coached weekly for two years and passed over for promotion twice. What was missing?
Coaching builds the capability; sponsorship makes it count. Someone who needs sponsorship and is offered coaching will feel well supported and go nowhere - and they cannot ask you for it directly, which is why it distributes so unevenly.